Debug: check runner WireGuard capability instead of UDP ports
Build Hugo / build (push) Failing after 13m45s
Build Hugo / deploy (push) Skipped

Other Netbird peers connect fine, so the server and 51820/udp are OK --
the hang is specific to the Gitea runner. Replace the UDP port probes
with runner-side checks: /dev/net/tun presence, effective capabilities
(NET_ADMIN), WireGuard tooling/module, and a definitive 'ip tuntap add'
test. A runner container lacking TUN/NET_ADMIN cannot create the
WireGuard interface, which explains 'netbird up' timing out after the
control plane connects.
This commit is contained in:
2026-07-25 01:53:09 +01:00
parent b05814980f
commit eacf71d676
+35 -28
View File
@@ -71,41 +71,48 @@ jobs:
echo "=== DNS resolution ==="
getent hosts "$host" || nslookup "$host" || echo "DNS resolution FAILED for $host"
# This deployment multiplexes management + signal + dashboard
# behind Traefik on :443 (gRPC-over-HTTP/2). There is no separate
# :10000 signal port. The data plane is WireGuard on 51820/udp and
# STUN on 3478/udp -- those are the ports that actually matter for
# 'netbird up' to complete the tunnel handshake.
echo "=== Control plane: HTTPS :443 (mgmt + signal via Traefik) ==="
curl -sS -o /dev/null -w "HTTP %{http_code} connect=%{time_connect}s total=%{time_total}s\n" \
echo "=== Control plane: HTTPS :443 reachable? ==="
curl -sS -o /dev/null -w "HTTP %{http_code} connect=%{time_connect}s\n" \
--max-time 15 "https://${host}/" || echo "curl to :443 FAILED"
ip="$(getent hosts "$host" | awk '{print $1; exit}')"
echo "Resolved IP: ${ip:-<none>}"
# Other Netbird peers connect fine, so the server + 51820/udp are OK.
# 'netbird up' still hangs on THIS runner, which points to the runner
# container being unable to create the WireGuard tunnel interface.
# WireGuard needs a TUN device + NET_ADMIN capability inside the
# container; runner containers often lack these.
echo "=== Data plane: WireGuard 51820/udp (nc UDP probe) ==="
if command -v nc >/dev/null 2>&1; then
nc -zu -w 5 "$host" 51820 \
&& echo "UDP 51820 appears OPEN/reachable" \
|| echo "UDP 51820 probe returned non-zero (may be filtered/blocked -- prime suspect)"
echo "=== Runner: TUN device present? (/dev/net/tun) ==="
if [ -c /dev/net/tun ]; then
echo "/dev/net/tun EXISTS (char device)"
else
echo "nc not available; installing..."
(apt-get update -qq && apt-get install -y -qq netcat-openbsd) >/dev/null 2>&1 || true
nc -zu -w 5 "$host" 51820 \
&& echo "UDP 51820 appears OPEN/reachable" \
|| echo "UDP 51820 probe returned non-zero (may be filtered/blocked -- prime suspect)"
echo "/dev/net/tun MISSING -- WireGuard cannot create a tunnel (LIKELY CAUSE)"
fi
echo "=== STUN 3478/udp (nc UDP probe) ==="
nc -zu -w 5 "$host" 3478 \
&& echo "UDP 3478 appears OPEN/reachable" \
|| echo "UDP 3478 probe returned non-zero (may be filtered/blocked)"
echo "=== Runner: effective capabilities (need NET_ADMIN) ==="
if command -v capsh >/dev/null 2>&1; then
capsh --print 2>/dev/null | grep -i 'Current:' || true
else
grep -i CapEff /proc/self/status || true
echo "(decode with: capsh --decode=<hex>; NET_ADMIN is bit 12 / 0x1000)"
fi
echo "=== Outbound UDP allowed at all? (probe public STUN) ==="
nc -zu -w 5 stun.l.google.com 19302 \
&& echo "General outbound UDP works (reached Google STUN)" \
|| echo "General outbound UDP may be BLOCKED on this runner network"
echo "=== Runner: WireGuard tooling / kernel module ==="
command -v wg >/dev/null 2>&1 && echo "wg present" || echo "wg NOT present"
command -v wireguard-go >/dev/null 2>&1 && echo "wireguard-go (userspace) present" || echo "wireguard-go NOT present"
(lsmod 2>/dev/null | grep -q wireguard && echo "kernel wireguard module loaded") || echo "kernel wireguard module NOT loaded (userspace fallback needed)"
echo "=== Runner: can we create a dummy TUN? (definitive test) ==="
if command -v ip >/dev/null 2>&1; then
if ip tuntap add dev nbtest0 mode tun 2>/tmp/tunerr; then
echo "SUCCESS: created TUN interface -> NET_ADMIN + TUN available"
ip tuntap del dev nbtest0 mode tun 2>/dev/null || true
else
echo "FAILED to create TUN interface -> $(cat /tmp/tunerr)"
echo "This is almost certainly why 'netbird up' hangs."
fi
else
echo "iproute2 'ip' not available"
fi
echo "=== done (this step never fails the build) ==="
continue-on-error: true